The Short Version

Most PMF survey templates are bloated. They ask 30 questions because "more data" feels smarter. It isn't. The Sean Ellis test works because it asks one question. The Vohra PMF Engine works because it builds a full measurement framework around four questions.

This guide gives you exactly those four questions, explains why each one works, flags the 20 questions you should delete, and includes a copy-paste template you can use today.

If you want to skip the explanation and grab the template, scroll to the bottom.

Why Most PMF Surveys Fail

The typical founder PMF survey looks like this: NPS question, then 15 questions about feature satisfaction, 5 questions about ease of use, a couple about pricing, and a free-text field at the end that nobody fills out.

This survey tells you nothing useful.

Why? Because it's a general feedback form dressed up as a PMF survey. It generates noise — a lot of noise — without giving you the signal you actually need: whether your users would miss your product if it were gone.

The four questions in this guide do one thing: they generate the data you need to make confident product and roadmap decisions. That's it. Every question earns its place.

The Four Questions That Actually Matter

Question 1: The Sean Ellis Core

"How would you feel if you could no longer use [product name]?"

Answer options:

Very disappointed
Somewhat disappointed
Not disappointed
N/A — I no longer use [product name]

Why this is the only question that matters first

Sean Ellis, who ran growth at Dropbox and HubSpot, discovered that the percentage of users who answer "very disappointed" predicts retention more accurately than any NPS score. Products where 40%+ of users answer "very disappointed" consistently showed stronger retention and growth. Products below that threshold consistently struggled.

This is your PMF score. It's a single number. Track it over time and you know whether you're getting stronger or weaker.

The other three options matter too — "somewhat disappointed" tells you you have something worth improving, "not disappointed" tells you users can live without you, and the N/A option filters out churned users who shouldn't be in your denominator.

One thing to note: don't change the wording. "How would you feel if you could no longer use Product X?" is the exact phrasing that produced the 40% benchmark. Changing it to something like "How important is Product X to you?" produces different results that don't map to the benchmark.

Question 2: The HXC Identifier

"What's the primary problem you were trying to solve when you started using [product name]?"

Answer options: (free text)

Why this question identifies your best customers

The high-expectation customers (HXC) — the ones who answered "very disappointed" — are the users whose feedback should drive your roadmap. Question 2 surfaces what problem they came to solve. When you cross-reference this with their "very disappointed" answer, you know exactly which use case is driving your strongest product-market fit.

Example: You run the survey. Your PMF score is 34%. Users who answered "very disappointed" and said "managing client projects" in the free-text field are your core HXC segment. That's the use case to double down on. The other segments — the ones where users are "somewhat disappointed" or "not disappointed" — are where you either improve or accept that the fit is weaker.

FitSignal runs this segmentation automatically. In a spreadsheet, you'd do this by filtering "very disappointed" responses and reading the free-text answers manually. Either way, you need this question.

Question 3: The Roadmap Splitter

"Which of the following best describes what you'd most like us to improve or add?"

Answer options:

[Specific feature your power users request most]
[Second most-requested feature]
[Third most-requested feature]
More use cases / broader functionality
Something else (free text)

Why this question decides your roadmap split

The Vohra PMF Engine recommends splitting your roadmap 50/50: 50% of engineering time on the features your HXC users are asking for, 50% on the features that might bring the "somewhat disappointed" users closer to PMF.

Question 3 tells you which bucket to put each feature request in. When a "very disappointed" user asks for something, it goes in the HXC bucket. When a "not disappointed" user asks for something, it goes in the expansion bucket (or gets deprioritized entirely).

Without this structure, you end up building based on whoever screams loudest. That's not a roadmap — that's noise.

Question 4: The Objection Surface

"What's the one thing that almost stopped you from using [product name]?"

Answer options: (free text)

Why this question reveals your biggest growth blocker

This is the question that surfaces the objection your best users almost didn't get past. For HXC users — the "very disappointed" segment — their answer tells you what they had to get over to become power users. Fix that objection for everyone, and your PMF score improves.

Example: Multiple HXC users answer "I almost didn't sign up because the onboarding seemed complicated." That's a signal. Improve the onboarding, and you remove the friction that stops more users from becoming HXC.

If you're using a spreadsheet, read the "very disappointed" responses to this question manually. In FitSignal, this is surfaced as part of the HXC user profile.

The 20 Questions to Delete

These are real questions I've seen in founder PMF surveys. None of them help you measure or improve PMF. Delete them:

How likely are you to recommend [product] to a friend? (NPS — not PMF)
How satisfied are you with [product]? (CSAT — not PMF)
How easy was [product] to use? (SUSt — not PMF)
Rate your overall experience (vague, no action)
What features do you use most? (useful analytics, not PMF data)
What features do you rarely or never use? (feature data, not PMF data)
How does [product] compare to competitors? (competitive intel, not PMF data)
What would make [product] better? (free-text fishing)
How long have you been using [product]? (account data, already in your CRM)
What is your role/title? (account data, already in your CRM)
How large is your company/team? (account data, already in your CRM)
How often do you use [product]? (engagement data, not PMF data)
What is your pricing plan? (account data, already in your CRM)
How did you hear about [product]? (acquisition data, not PMF data)
Would you pay more for [product]? (pricing research, not PMF data)
What integrations do you need? (product roadmap input, not PMF data)
How satisfied are you with our support team? (support CSAT, not PMF)
Did [specific feature] solve your problem? (single-feature feedback, not PMF)
How likely are you to continue using [product]? (retention intent, conflated with PMF)
What could we do to surprise and delight you? (marketing language, no actionable signal)

The pattern: most of these questions answer useful business questions — but they're not PMF questions. Run them as separate surveys if you need that data. Don't混 them into the PMF survey where they dilute your signal.

Survey Timing: When to Ask

You can't ask a new user the "very disappointed" question on day one. They don't know yet whether they'd miss something they just started using.

The standard rule: ask users who've been active for at least 2 weeks and have completed at least 3 meaningful interactions with your product. If you have a freemium or trial product, wait until they've hit whatever activation milestone correlates with long-term retention.

Minimum sample size: 30 responses before your score is meaningful. Below 30, the margin of error is too wide. 50+ is better. If you're a small product with under 100 users, survey everyone who qualifies — don't sample.

Survey Cadence: How Often to Run It

PMF drift is real. You can have PMF and lose it. Instagram's PMF weakened after Reels. Netflix's weakened as streaming became commoditized.

The minimum: quarterly. Run the same four questions every 90 days to the same user cohort. Track whether your "very disappointed" percentage is moving.

Better: monthly, if you have the volume to make the data meaningful. At 50+ responses per run, monthly gives you a faster feedback loop on product changes.

Copy-Paste PMF Survey Template

Copy the text below directly into your survey tool (Typeform, Google Forms, Survicate — any of them):

Q1: How would you feel if you could no longer use [Product Name]?

Very disappointed
Somewhat disappointed
Not disappointed
N/A — I no longer use [Product Name]

Q2: What primary problem were you trying to solve when you started using [Product Name]? [Free text]

Q3: Which would you most like us to improve or add?

[Feature A — your top requested feature from power users]
[Feature B]
[Feature C]
More use cases / broader functionality
Something else [free text]

Q4: What almost stopped you from using [Product Name]? [Free text]

Q5 (optional): What's your email? [Free text — only if you want to follow up with respondents]

How to Calculate Your Score

Your PMF score = (Number of "Very Disappointed" responses ÷ Number of valid responses) × 100

Valid responses = all responses except "N/A — I no longer use [Product Name]"

Example: 100 valid responses, 38 "Very Disappointed" = 38% PMF score

Benchmark: 40%+ = strong PMF. Below 40% = keep iterating.

What to Do With the Data

Once you have your score and your HXC segments, the Vohra PMF Engine says:

Survey — you just did this
Segment — identify your "very disappointed" users and cross-reference with Q2 (the problem they came to solve)
Analyze — what do your HXC users have in common? What's the objection they had to get past (Q4)?
Implement — fix the objection for the broadest segment; add the feature your HXC users are requesting
Track — run the same survey in 90 days and see if your score improved

FitSignal runs this loop automatically. In a spreadsheet, you track it manually.

The Bottom Line

Four questions. That's all you need.

The Sean Ellis question gives you your score. Question 2 tells you who your best users are and what they came for. Question 3 splits your roadmap. Question 4 tells you what to fix.

Everything else is noise that makes your survey harder to analyze and your users less likely to complete it.

Run these four questions to your active, qualified users every quarter. Track your score. Improve the objections. Add what your HXC users need. Run it again.

Try FitSignal — auto-scores your PMF, auto-segments by HXC, tracks your trend over time →